The Stanford Coronary Risk Intervention Project was a four-year randomized clinical trial that showed that risk reduction through lifestyle change and lipid-lowering medications significantly reduced the rate of narrowing of the minimum diameter of coronary artery segments with angiographically visible lesions in 119 patients vs. 127 controls who received usual physician care. In collaboration with this trial, Dr. Ronald Krauss measured high-density lipoprotein (HDL) subclasses by gradient gel electrophoresis. HDL may be divided into two HDL2 and three HDL3 subclasses that are approximated by their estimated particle diameters: HDL3c (7.2-7.8 nm), HDL3b (7.8-8.2 nm), HDL3a (8.2- 8.8 nm), HDL2a (8.8-9.7 nm) and HDL2b (9.7-12.9 nm). The HDL- distribution can also be characterized by the diameter of the predominant peak, which may lie in either the HDL3b or HDL3a interval. Case control and angiographic studies suggest that coronary heart disease risk is increased when HDL2b is reduced relative to HDL3c and HDL3b Our objective is to assess the influence of HDL-subclasses with coronary disease progression, and to identify factors influencing HDL subclasses at baseline and over time. The specific questions to be examined are: 1. Did the risk reduction program change specific HDL subclasses as compared to controls? 2. Did the HDL gradient gel profile characterize men most likely to benefit from multifactor risk reduction? 3. Did HDL-subclasses change significantly in patients that reduced fat intake, reduce body weight, or who took one or more of the following medications: colestipol, nicotinic acid, clofibrate, probucol, gemfibrozil, fenofibrate, lovastatin, guar gum or fish oils? 4. What are the cross-sectional associations of HDL-subclasses with adiposity, fasting and post-load insulin and glucose, diet and medications at baseline? Preliminary analyses suggest that: 1) During the trial, men in the treatment group increased HDL2b; 2) the special intervention was most effective in reducing coronary disease progression in subjects with a baseline predominant HDL-peak diameter below the median; 3) HDL- subclasses were more strongly influenced by diet and adiposity than by drugs during the trial; 4) carbohydrates, alcohol and caffeine were associated with specific subclasses at baseline. The data are stored in computer files, and except for the analyses contained in this application, are unanalyzed. These data were collected as part of NIH grant Plasma lipoproteins in coronary artery disease, HL-33577.